Analysis of a capacitance tomography system for object detection and localization

To ensure safe human-robot cooperation, it is necessary not only to measure the force exerted during contact, but also to monitor an approaching action. This would allow avoiding any unintentional contact through evasive maneuvers and thus prevent possible injuries. In our contribution we analyze a localization system based on capacitance tomography where an approaching object constitutes an electrode of a multi-electrode capacitance. The system processes measured voltages made on the surface of a conductive layer. From this, the location of the capacitive coupling between the object and the layer can be reconstructed, and the position of the object can be estimated. The system's performance is further analyzed using the Neyman-Pearson detector. To define the performance of the system, for object detection only, the influence of object distance, object area, and number of measurements used on the probability of detection is examined.